JPH1048246A - Semiconductor acceleration sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a small high sensitivity sensor by forming a cantilever or a bridge, as a mass for movable electrode, of a metal silicide layer having high density. SOLUTION: A silicon oxide is deposited on the major surface of a semiconductor substrate 11 and an opening is made in a specified region. A silicon single crystal epitaxial layer is then grown selectively at the opening and an amorphous silicon layer is grown on the entire surface. Tantalum is then deposited on the amorphous silicon layer and patterned into a specified shape and then a fixed electrode 12 and a cantilever movable electrode 13 having one fixed to the semiconductor substrate 11 are formed before being annealed to form a high density tantalum silicide layer. Subsequently, underlying silicon oxide is removed by etching so that the movable electrode 13 of silicide layer can be displaced. Finally, the fixed electrode 12 and the movable electrode 13 are connected to form a capacitor bridge. Since the movable electrode 13 is formed of a high density metal silicide layer, output voltage from the bridge can be increased.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor acceleration sensor, and more particularly, to a miniaturized and high-sensitivity acceleration sensor for an airbag system and a suspension control system for an automobile.

[0002]

2. Description of the Related Art In recent years, an apparatus has been proposed in which a small acceleration sensor is formed on the surface of a silicon semiconductor substrate by using micromachining technology, and a circuit for processing an operation signal is integrated and integrated on the substrate. I have.

[0003] In these, a mass for a movable electrode such as a cantilever (cantilever) or a bridge of an acceleration sensor is formed of a silicon single crystal layer or a silicon polycrystal layer.
In the case of a silicon single crystal layer, a silicon-on-insulator (SOI) substrate is used. In the case of a silicon polycrystalline layer, it is formed by the same CVD process as that for manufacturing a general silicon semiconductor device.

[0004] Related devices of this type include, for example, JP-A-5-304303, JP-A-4-504003, JP-A-62-123361, U.S. Pat.
And the like.

[0005]

In the above prior art, a silicon layer is used as the material of the mass for the movable electrode. Silicon has a relatively low density compared to other metals.
Therefore, it is difficult to achieve both high sensitivity and miniaturization.

In a capacitive acceleration sensor, the amount of change in capacitance is measured from the displacement of a mass that is a movable electrode.
At this time, the output voltage from the capacitance increases in proportion to the mass of the movable electrode, that is, the sensitivity can be increased. Alternatively, in order to achieve high sensitivity, it is necessary to increase the mass of the movable electrode, that is, to increase the size of the movable electrode, and it is difficult to achieve both reduction in size and cost.

An object of the present invention is to provide a low-cost acceleration sensor that is small, has high sensitivity, is easy to manufacture, and has a low cost.

[0008]

The above object can be achieved by forming a cantilever (cantilever), bridge, or the like as a mass for a movable electrode with a metal silicide layer having a high density.

[0009]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a partial perspective view of a capacitance detecting section of an acceleration sensor according to a first embodiment of the present invention, and FIG. 2 is a cross-sectional view taken along the line AA 'in FIG. is there.

In FIG. 1, a fixed electrode 12 and a movable electrode 13 of a cantilever having one end fixed to the substrate are formed on an upper surface of a silicon semiconductor substrate 11, and the respective electrodes are connected to form a bridge of a capacitor. ing. Although not shown in the drawing, a signal processing circuit is formed at a predetermined position on the silicon semiconductor substrate 11, and the fixed electrode 1
2 and an output terminal from the movable electrode 13.

Next, a method of manufacturing the semiconductor acceleration sensor according to the present invention will be described.

FIG. 2A shows a silicon semiconductor single crystal substrate 2.
1 is shown. The quality of this crystal is as follows: production method: CZ (Czochralski method; pulling method), crystal orientation; (100) plane with 3 ± 0.5 ° off-angle (off-orientation),
Conductivity type; n-type; resistivity; 10 Ω-cm. The off-angle is for preventing the occurrence of stacking faults in the epitaxial process. Although not shown in the drawing, a signal processing circuit is formed in a predetermined area by using a normal CMOS process or a bipolar process.

FIG. 2B shows a silicon semiconductor single crystal substrate 2.
1 shows a state in which a silicon oxide film 22 is formed on the main surface and an opening 23 is provided in a predetermined region. Silicon oxide film 22
Has a thickness of 1-2 μm and is formed by thermal oxidation or CVD.
Method can be used. Note that a stacked film of a silicon nitride film and a silicon oxide film can be used instead of the silicon oxide film 22.

FIG. 2C shows a state in which a silicon single crystal epitaxial layer 24a is selectively grown in the opening and an amorphous or polycrystalline silicon layer 24b is grown on the entire surface. The silicon single crystal epitaxial layer 24a has a substrate temperature of 78
At 5 ° C., silane gas (SiH ₄ ) is used as a raw material.
It is grown by adding 0.1% hydrogen chloride gas. The thickness is the same as the thickness of the silicon oxide film 22, that is, 1 to 2 μm.
It is. This is because the substrate temperature during the growth does not harm the signal processing circuit. Even though it is a single crystal due to the low temperature, it has many dislocations and stacking faults as compared with the epitaxial layer used for a general LSI, but the quality is inferior. However, there is no particular obstacle in using the present invention. Further continuously, an amorphous or polycrystalline silicon layer 24b is formed over the entire surface of the substrate.
Was deposited. This is a process in which the substrate temperature is set to 550 ° C., the addition of hydrogen chloride gas is stopped in the atmosphere, and the entire surface of the substrate is subjected to a CVD reaction in the same reaction apparatus following this step.

FIG. 2D shows a state in which a tantalum film 25 is deposited and patterned on the amorphous or polycrystalline silicon layer 24b. The tantalum film 25 is formed by sputtering and has a thickness of 2.5 to 3.0 μm. The shape of the pattern is a region where the movable electrode and the fixed electrode are formed. Patterning is performed by a usual photolithography method and a wet etching method using a mixed solution of hydrochloric acid and nitric acid.

FIG. 2E shows a state in which the amorphous or polycrystalline silicon layer 24b is patterned into a predetermined shape. The shape of the pattern is fixed electrode 1 as shown in FIG.
The fixed end for 2 and the mass for the movable electrode 13 are formed separately. Amorphous or polycrystalline silicon layer 24b
Is performed by anisotropic processing using ordinary photolithography and microwave-excited plasma etching.

FIG. 2F shows a tantalum film 2 formed by annealing.
5 shows a state in which a tantalum silicide layer 26 is formed by reacting No. 5 with an amorphous or polycrystalline silicon layer 24b. The annealing condition is a temperature of 600 to 7 in an atmosphere of a helium gas stream or a mixed gas stream of helium and hydrogen.
50 ° C., time 30 minutes to 2 hours. At this time, tantalum having a thickness of 1 unit reacts with silicon having a thickness of about 2.2 units to form a tantalum silicide layer having a thickness of about 2.4 units.
If the tantalum film is excessively thick, a two-layer structure in which an unreacted tantalum film remains on the silicide layer, and if the silicon layer is excessively thick, a two-layer structure in which an unreacted silicon layer remains under the silicide layer. . In each case, the layer structure and its thickness can be controlled by adjusting the respective thicknesses. In the silicidation reaction, volume shrinkage occurs, and tensile stress is generated in the film. However, the stress can be reduced by forming a two-layer laminated film structure in which the silicon layer 24 is partially left. Further, in order to reduce the stress generated due to the difference in the coefficient of thermal expansion caused by the temperature during the reaction, it is necessary to precisely control the annealing temperature, and for tantalum silicide, around 670 ° C. is suitable.

At the time of this annealing, the amorphous or polycrystalline silicon layer 24 below the tantalum film and other regions is used.
b is also annealed at the same time to promote crystallization, and solid-phase growth starts from the portion in contact with the silicon single-crystal epitaxial layer 24a to form the high-quality silicon layer 24 oriented in the same crystal orientation as the silicon single-crystal epitaxial layer 24a.
Here, the high-quality silicon layer means a layer having a large crystal grain size and excellent crystal orientation. This is very effective in preventing the bending and deformation of the silicon layer of the beam portion 27 that connects and suspends the movable electrode mass with the fixed end, and achieves high reliability.

The density of the tantalum silicide thus formed is about 2.53 times that of silicon when calculated from the crystal structure and the molecular density.

FIG. 2G shows a state in which the underlying silicon oxide film 22 is removed by etching. The etching of the silicon oxide film 22 is performed by immersing in a mixed solution of hydrofluoric acid and ammonium fluoride and removing not only the surface exposed portion but also the lower portion of the movable electrode 26 and the lower portion of the beam portion 27 connected to the fixed end by side etching. It is.

As a result, the silicide layer 26 can be movably displaced.

As the metal silicide layer, a material satisfying the following conditions can be employed other than tantalum.

(1) High density (preferably about twice or more the density of silicon).

(2) Being compatible with semiconductor processes (manufacturing temperature, chemical resistance, not deteriorating the properties of silicon or silicon oxide film, etc.).

(3) Mechanically stable (adhesion, stress, etc.).

(4) No change over time.

(5) The manufacturing is easy and the cost is low.

Such a metal material is represented by IVb in the periodic table of the elements.
(Titanium, zirconium, halfnium), Vb (vanadium, niobium, tantalum) and VIb (chromium, molybdenum, tungsten) group metals. In particular, the disilicide of these metals forms a silicide layer with almost the same volume as the original silicon layer by the solid phase reaction, so that the surface is kept almost flat, and by appropriately selecting the temperature of the solid phase reaction It is excellent because the internal stress can be reduced and its change with time can be reduced.

[Second Embodiment] Next, the second embodiment will be described with reference to the first embodiment.
The following description will focus on the differences from the embodiment.

In the first embodiment, both the fixed electrode and the movable electrode are formed from a silicon layer deposited on a silicon semiconductor substrate. However, the fixed electrode is previously placed at a predetermined position on the silicon semiconductor substrate in a diffusion layer having a predetermined pattern. It is also possible by forming.

FIG. 3 is a partial perspective view of a capacitance detecting section of an acceleration sensor according to a second embodiment of the present invention, and FIG. 4 is a sectional view of the AA 'section in FIG.

In FIG. 3, a movable electrode mass body 33 is formed of a fixed electrode in the main surface of the upper surface of the silicon semiconductor substrate 31 and a silicon layer deposited on the main surface of the silicon semiconductor substrate 31. The movable electrode mass 33 is connected to the fixed end 35 by beam portions 34 hanging from the four corners. Each electrode is connected to form a capacitor bridge. Although omitted in the drawing, a signal processing circuit is formed at a predetermined position on the silicon semiconductor substrate 31,
It is connected to output terminals from the fixed electrode and the movable electrode.

Next, a method of manufacturing the semiconductor acceleration sensor according to the present invention will be described.

FIG. 4A shows a silicon semiconductor single crystal substrate 4.
1 is shown. The quality of this crystal is the same as that of the first embodiment.

FIG. 4B shows a silicon semiconductor single crystal substrate 4.
1. Boron is diffused into a region where a fixed electrode is
8 is formed, a silicon oxide film 42 is formed thereon, and an opening 43 is provided in a predetermined region. Although the formation of the p-type layer 48 is omitted in the drawing, it can be formed simultaneously with the formation of the signal processing circuit.

FIG. 4C shows a silicon single crystal epitaxial layer 44a and an amorphous or polycrystalline silicon layer 44.
This shows a state where b has grown. This is similar to the first embodiment.

FIG. 4D shows a state in which a tantalum film 45 is deposited and patterned on the amorphous or polycrystalline silicon layer 44b. The shape of the pattern is the center of the region that will be the mass of the movable electrode. The thickness of the tantalum film 45 is 3 μm, and the amorphous or polycrystalline silicon layer 44 b
In the thickness direction.

FIG. 4E shows a state where the amorphous or polycrystalline silicon layer 44b is patterned into a predetermined shape. The shape of the pattern is X in the plane as shown in FIG.
A movable electrode in the form of a comb extending in the two-dimensional direction Y and a suspension beam are formed separately.

FIG. 4F shows a tantalum film 4 formed by annealing.
5 shows a state in which a tantalum silicide layer 46 is formed by reacting No. 5 with an amorphous or polycrystalline silicon layer 44b.

The annealing conditions are the same as in the first embodiment.

In the first embodiment, the mass of the movable electrode has a structure in which a metal silicide layer and a silicon layer are vertically stacked. Here, the mass of the movable electrode is formed on both sides of the metal silicide layer. This is a three-row structure in which layers are left.
The reason why not only the silicide layer but also the silicon layer is left and used is to relieve stress as in the case of the first embodiment.

If an excess unreacted tantalum film remains, it is removed by etching with nitric acid or the like.

FIG. 4G shows a state in which the underlying silicon oxide film 42 is removed by etching. The etching of the silicon oxide film 42 is performed by immersing in a mixed solution of hydrofluoric acid and ammonium fluoride and removing not only the surface exposed portion but also the lower portion of the beam portion connected to and suspended from the movable electrode and the fixed end by side etching. .

In the embodiment of the present invention, a direct solid-phase reaction between metal and silicon is used as a method for forming a silicide layer. However, other methods, such as simultaneous sputtering or vapor deposition of metal and silicon, and metal silicide sputtering, It is also possible to use CVD of metal silicide.

[0046]

According to the present invention, the output voltage from the capacitance can be increased by forming a movable electrode mass such as a cantilever or a bridge from a high-density metal silicide layer. Alternatively, even if the size of the movable electrode mass is reduced to half or less, an output voltage equivalent to that of the related art can be obtained.
As a result, the acceleration sensor can be reduced in size and sensitivity.

[Brief description of the drawings]

FIG. 1 is a perspective view of a main part of a sensor according to a first embodiment of the present invention.

FIG. 2 is a sectional view taken along the line AA ′ of FIG. 1 for illustrating the manufacturing process of the first embodiment of the present invention.

FIG. 3 is a perspective view of a main part of a sensor according to a second embodiment of the present invention.

FIG. 4 is a sectional view taken along the line AA ′ of FIG. 1 for illustrating a manufacturing process according to a second embodiment of the present invention.

[Explanation of symbols]

11, 21, 31, 41 ... a silicon single crystal semiconductor substrate,
12: fixed electrode portion, 13, 33: mass of movable electrode, 2
2, 42 ... silicon oxide film, 24, 44 ... silicon layer,
25, 45 ... metal film, 26, 46 ... metal silicide layer,
27, 34: suspended beam part, 35: fixed end, 48: diffusion layer.

Claims (2)

[Claims] 1. A capacitive sensor for detecting acceleration formed on a silicon semiconductor substrate, comprising: at least one fixed electrode; and a movable electrode movable in response to the acceleration facing the fixed electrode. A mass of the movable electrode is formed of a layer mainly composed of metal silicide. 2. The acceleration sensor according to claim 1, wherein the metal of the metal silicide layer is one or more of metals belonging to IVb, Vb, and VIb of the periodic table.